Child sexual abuse in India, in numbers.

Three studies get quoted most often. They asked different people different questions, so their answers cannot be set side by side as one number.

What all three agree on is that sexual abuse of children in India is common, that it reaches both girls and boys, and that most of it is never reported. Police records, below, count something else: the cases that reached a police station.

Kaaval’s essay on how a child actually tells starts from that silence.

The study still matters. It found sexual abuse reported by girls and boys across the states and settings it covered, and most of those children had told no one. It cannot give the share of all Indian children. For that, the Lancet estimate in the table above is the most complete current guide, with its wide margins.

NCRB counts cases, which means First Information Reports (FIRs), not children. One FIR can name more than one child: the 69,191 cases in 2024 involved 70,132 children. When an FIR lists several offences, NCRB counts it once, under the most serious one.

POCSO cases were 36.9 percent of the 1,87,702 cases of crime against children police registered in 2024.

The count has risen every year but one since 2017. NCRB itself warns that a rise in crime and a rise in registered crime “are clearly two different things”. More awareness, the 1098 helpline’s link with 112, online complaints and readier police can all bring more cases in. The numbers cannot show how much of the rise is more abuse.

The 96.6 percent covers penetrative sexual assault cases under sections 4 and 6 of the POCSO Act, 44,126 cases in 2024, and it counts cases, not children. Half of those cases, 22,308, sit in a column NCRB labels “friends, online friends or live-in partners on pretext of marriage”. More than half the children in all these cases, 52.7 percent, were 16 or 17. Some of these are likely to be teenage relationships that a family reported to the police. Others are adults who used an online friendship or a promise of marriage to abuse a teenager. In Mumbai, lawyers at Majlis, a legal aid organisation, say that in the cases they see, the other person is often much older, and the child is sometimes trying to leave a violent home. NCRB does not separate them, so no one can say how many.

The family column, 3,658 cases or 8.3 percent of the 2024 total, probably undercounts abuse at home. Families are often the last to report one of their own. Read when it is someone in the family.

In India, the law treats anyone under 18 as a child. Any sexual contact with someone under 18 is an offence, even if both are teenagers and both agreed. Courts have asked the government to consider changing this for teenagers close in age. The law has not changed.

Most children who are harmed know the person who harms them: a relative, a neighbour, a teacher, a friend. Set aside every case in the friends and partners column, and the child still knew the accused in 93 percent of the rest.

For who that circle takes in, read the ring you have not been counting, and when an adult’s behaviour is not okay.

Protection, medical care and support can begin within days of a report. A support person is a person assigned by the Child Welfare Committee and paid by the government, who stays with your family through the police and court steps. In practice, support is thin. A 2025 review by Maharashtra’s child rights commission and Majlis found that in 2021 and 2022, fewer than 1 in 10 cases had a support person, and state compensation reached fewer than 3 in 10. Read what happens after a report.

Most cases do reach a court. In 2024 police sent 65,113 POCSO cases for trial. That is 95.3 percent of the cases police closed that year, NCRB’s charge-sheeting rate; 21,256 were still under investigation at year end. The wait comes after that.

The law asks the Special Court to finish a trial within a year of taking up the case, “as far as possible”. At the end of 2024, 2,86,903 POCSO cases were waiting for trial in India. Courts finished 40,299 that year, 1,050 of them without a trial. At that pace, courts would need about seven years to clear the backlog even if no new case came in. That is not how long one case waits. NCRB does not publish how long each case waits. The number waiting grew by 24,814 in 2024, because more cases came in than courts could finish.

Of the 39,249 trials completed in 2024, 12,672 ended in conviction, 25,476 in acquittal and 1,101 in discharge. An acquittal means the court did not find the charge proved to the standard the law requires. It does not, on its own, mean that nothing happened.

Kerala registered 4,513 POCSO cases in 2024, a rate of 48.0 for every lakh children. Bihar, with far more children, registered 2,468, a rate of 5.2. Tamil Nadu registered 6,969, a rate of 33.7, and Uttar Pradesh 8,087, a rate of 9.5.

A high rate can mean that children tell, families report and police register. A low rate can mean the opposite. NCRB prints its own warning on its state tables: states “may not be compared purely on the basis of crime figures”.

There is a second problem. NCRB works out every rate using the number of children counted in the 2011 Census, now given as 44.4 crore, and has not updated it since. NCRB uses yearly population projections for its overall and women’s crime rates, but the 2011 count for children. Where the number of children has fallen since 2011, the rate is understated. Where it has grown, the rate is overstated.

In police records, boys appear rarely. Of those 892 boys, 394 were under 12. The number of boys was 423 in 2023, so it more than doubled in a year. Most of the rise came from Kerala and Maharashtra. NCRB does not say why.

Surveys find far more. In the 2007 government study, boys reported sexual abuse about as often as girls.

Read Still a Boy.

The two numbers measure different things. NCMEC prints its own warning: the reports “are not indicative of the level of child sexual abuse in a particular country”.

If you have come across such material, here is where to report it online. On children’s photos and fakes, read The Dot Behind the Ear.